The structure of the American family has changed significantly -- particularly in terms of size and parental presence. Objective data are needed about actual behavior within the family unit. A conceptual model is proposed to account for the impact of structural variables on the functioning of individual families. General structural variables are hypothesized to exert a more predictable impact on a family when parents' decision making strategies are poor, when positive feelings and commitment within the family are low, and when external social support is minimal. Families (300 black, 300 white/non-Hispanic) will be selected on the basis of number of children (1, 2, and 4 or more) and parental presence (1 vs 2). Family functioning will be assessed via systematic home-based observations, standardized interviews, environmental evaluations, self-ratings, and activity logs. Patterns of social interaction will be coded systematically during 3 4-hr home visits, focusing on amounts and types of interaction among all family members, as well as expression of feelings, household activities, discipline, and teaching. Parents' decision making strategies will be evaluated formally (3 sessions) concerning family decisions related to structure, childcare, discipline, and division of responsibilities. Parents' perceptions of the advantages and disadvantages associated with their own family situation will be measured, as will their current satisfaction and strategies used to solve problems. Child outcomes include school performance, scores on standardized tests, parent ratings of abilities and behavioral characteristics, and observed social and verbal behavior. Variables to be considered in data analysis and interpretation include: age, sex, and spacing of children; stability of family membership; religiosity; and parents' current age, marital history, education, and current work status. The findings are relevant to understanding why some families function well in spite of serious structural obstacles, while others do not. The theoretical model to be tested has significance for developing public policy and family related programs, as well as for increasing scientific knowledge about the impact of environmental variables on children's intellectual and social development.